Statalist The Stata Listserver


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: Xtivreg with categorical endogenous covariate


From   "Schaffer, Mark E" <M.E.Schaffer@hw.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: Xtivreg with categorical endogenous covariate
Date   Thu, 4 May 2006 15:02:50 +0100

Dimitris, 

> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu 
> [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of 
> Dimitris Pavlopoulos
> Sent: 04 May 2006 14:45
> To: statalist@hsphsun2.harvard.edu
> Subject: st: Xtivreg with categorical endogenous covariate
> 
> Dear all,
> 
> I want to run panel regression with a continuous dependent 
> (wage). Among the covariates I use, one (job mobility) is 
> endogenous. Therefore, I want to use a 2SLS fixed effects 
> regression. However, my endogenous covariate is categorical 
> (has 3 values 0,1,2). 
> 
> What I would like to know is whether I can use xtivreg, fe 
> for this estimation. Although not mentioned specifically in 
> the help for xtivreg, I think that this command can be 
> applied only if the endogenous covariate is continuous and 
> NOT when it is categorical.

This question comes up regularly on the list in the context of a dummy
endogenous variable in standard IV, e.g.,

http://newwww.stata.com/statalist/archive/2004-07/msg00710.html

The difference here is you have a categorical variable.  The answer is
that this is not an IV question, it's a functional form question.  If
you think that a linear specification is legitimate, then panel IV is
fine too.  If you think that a linear specification is not legitimate,
then endogeneity doesn't matter - neither OLS nor standard fixed effects
nor xtivreg (or xtivreg2) will be legitimate, and you'll need to do
something along the lines you're suggesting.

Hope this helps.

Cheers,
Mark
 
> 
> If xtivreg cannot be applied in my case, would you have any 
> other suggestions for doing the estimation procedure? For 
> example could I run an mlogit or an mprobit for the first 
> step and then plug the predicted probabilities in the primary 
> equation in replace for the endogenous covariate? But then 
> how can I get correct standard errors?
> 
> 
> Thank you in advance,
> Dimitris Pavlopoulos
> 
> 
> *
> *   For searches and help try:
> *   http://www.stata.com/support/faqs/res/findit.html
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/
> 
> 

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index